Last week, I hopped on a plane to Las Vegas for a two-night stay at Caesar’s Palace. But I wasn’t there for the gambling or the faux-Roman statues. I was attending the ABA’s 2026 Class Actions National Institute, which Taft sponsored. Here are some of my takeaways from the conference.   

1. AI and Class Actions. Of course, AI was a major topic at the conference. Nearly every panel discussed the effects of AI on class-action practice.
   1. Algorithmic price-fixing. The panel on algorithmic price-fixing class actions presented an interesting frame work for how to think about these cases: what information is shared, how is it used, and how much discretion do the users have in deciding whether to adopt the algorithm’s recommendations.
   2. AI model claims. A separate panel focused on claims related to AI models. It identified two developing trends. First, that claims against the AI models will progress from being brought primarily by authors and other writers of text to claims brought by musicians and video-content creators. And second, that claims would move from being brought against only the AI model companies to being brought against companies that use the AI models in their business.
2. The “Year in Review” Session. A regular highlight of this conference. Without any major Supreme Court cases related to class certification this year, the presenters, Myriam Gilles and Adam Zimmerman, focused on three areas where the lower courts are working out issues left open by earlier Supreme Court opinions:
   1. Standing. TransUnion LLC v. Ramirez, 594 U.S. 413 (2021), left open “whether every class member must demonstrate standing before a court certifies a class.” The Court had accepted LabCorp v. Davis, 605 U.S. 327 (2025) to resolve this issue but then dismissed it as improvidently granted. This year, the 4th, 9th, and 10th Circuits all issued opinions related to the standing question. The 9th Circuit’s decision in Healy v. Milliman, 164 F.4th 701, may be the most consequential. It requires that there must be sufficient evidence at summary judgment for absent class members to create a question of fact as to standing.
   2. Damages models. Under Comcast Corp. v. Behrend, 569 U.S. 27 (2013), plaintiffs must put forth, at the class-certification stage, evidentiary proof of a damages methodology showing how damages are calculable on a class-wide basis in a manner consistent with their theory of liability that is not speculative. This year, in Office of the Gen’l Treasurer v. Boeing Co., 2026 WL 2083048 (4th Cir. July 20, 2026), the 4th Circuit held that the plaintiff’s expert had not satisfied Comcast because he did not identify a specific methodology he would use to calculate the alleged damages but rather presented only approaches he could use.
   3. Personal jurisdiction. In Bristol Myers Squibb Co. v. Superior Court of Cal., 582 U.S. 255 (2017), a mass action involving hundreds of individual plaintiffs, the Court held that, unless there is general personal jurisdiction, each plaintiff in the case must be able to establish specific personal jurisdiction against the defendant. Justice Sotomayor’s dissent in that case asserted that the Court did not decide whether the rule also applies to class actions. Since then, a majority of circuit courts have held that Bristol Myers Squibb applies to FLSA collective actions, with the 2d, 3d, 6th, 7th, 8th, and 9th Circuit holding the majority position. The lone outlier is the 1st Circuit.

The National Institute is one of the best annual class-action conferences. The other panels were equally interesting, including ones discussing judicial scrutiny of settlements and the process related to matters before the Judicial Panel on Multidistrict Litigation. If it fits with my schedule next October, I hope to be taking the road to New Orleans for next year’s conference.